# Failed to Import transformers.models
**URL:**
**Author:**  [@Henrkkn](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2FHenrkkn)%5C
**Post date:** [June 15, 2023, 4:59am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Ffailed-to-import-transformers-models%2F43315%2F1 "2023-06-15T04:59:40Z")
Hello!
I need some help to fix my “RunTimeError” message.
The code used:
from transformers import TFAutoModel
checkpoint=“distilbert-base-uncased-finetuned-sst-2-english”
model=TFAutoModel.from\_pretrained(checkpoint)
**And the long error message**
RuntimeError: Failed to import transformers.models.distilbert.modeling\_tf\_distilbert because of the following error (look up to see its traceback):
No module named ‘keras.saving.hdf5\_format’
I cant see where I have done something wrong, it seems more there is something wrong with the model?
---
**Author:**  [@chem1](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fchem1)%5C
**Post date:** [June 15, 2023, 7:55am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Ffailed-to-import-transformers-models%2F43315%2F2 "2023-06-15T07:55:13Z")
> [@Henrkkn](#):
>
> No module named ‘keras.saving.hdf5\_format’
@Henrkkn Have you tried updating TensorFlow?
```auto
pip install -U tensorflow==2.10
```
---
**Author:**  [@Henrkkn](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2FHenrkkn)%5C
**Post date:** [June 15, 2023, 9:35am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Ffailed-to-import-transformers-models%2F43315%2F4 "2023-06-15T09:35:14Z")
After installed the Tensorflow as mentioned, I now receive a new RunTime Error:
"RuntimeError: Failed to import transformers.models.distilbert.modeling\_tf\_distilbert because of the following error (look up to see its traceback):
Duplicate registrations for type ‘experimentalOptimizer’ "
---
**Author:**  [@rifazks](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Frifazks)%5C
**Post date:** [September 15, 2023, 3:32am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Ffailed-to-import-transformers-models%2F43315%2F5 "2023-09-15T03:32:51Z")
Rolling back to 2.10 results out with google protobuf disconnect error.
---
**Author:**  [@Tarzan-07](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2FTarzan-07)%5C
**Post date:** [March 19, 2024, 6:55pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Ffailed-to-import-transformers-models%2F43315%2F6 "2024-03-19T18:55:59Z")
@chem1 I tried this:
from transformers import TFAutoModel
checkpoint = ‘distilbert-base-uncased-finetuned-sst-2-english’
model = TFAutoModel.from\_pretrained(checkpoint)
I am getting this:
RuntimeError: Failed to import transformers.models.distilbert.modeling\_tf\_distilbert because of the following error (look up to see its traceback):
module ‘tensorflow.\_api.v2.compat.v2. **internal** ’ has no attribute ‘register\_load\_context\_function’
I am using latest version of everything. More specifically, transformers = 4.38.2, tensorflow = 2.16, keras = 3.1.1, protobuf = 5.26.0
---
**Author:**  [@Sandy1857](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2FSandy1857)%5C
**Post date:** [March 20, 2024, 12:25am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Ffailed-to-import-transformers-models%2F43315%2F7 "2024-03-20T00:25:40Z")
Maybe the issue is because of keras 3 which is the default since tf 2.16. suggest trying out with lower tf version.